The GTM hunt: turn your knowledge base and live web search into a customer-finding machine

Most prospecting tools start with a filter: industry, headcount, geography. That produces a list that looks exactly like everyone else’s list. The GTM hunt starts somewhere better: your own knowledge base.

Connect Notion and point Walle at the pages that describe your customers. Meeting notes from sales calls. The ICP doc someone wrote in a burst of clarity last quarter. Win and loss notes. Onboarding retros. Walle reads them the way a sharp new hire would, and builds a working picture of who actually buys from you and why.

A knowledge base tells Walle who your best customers are. Web search lets it find the next ones. Connect Parallel and Tavily and Walle scours the live web instead of guessing from stale training data.

The two are complementary. Parallel is built for wide structured sweeps: find every company that matches a profile. Tavily is fast targeted lookup: verify a fact, pull a filing, check who runs ops at a specific firm. Walle picks the right tool for each step and stitches the results together.

Lookalike lists are table stakes. The interesting work is teaching Walle the signals that mark a company as ready right now. Tell it once, in plain language, and it hunts for them on every sweep:

• Hiring spikes in ops and compliance. A company posting five compliance analyst roles is drowning in manual process.
• Trigger events: an ERP migration, a new regulation landing, audit pressure, a funding round with an operations mandate attached.
• Document-heavy wedge processes. Onboarding packets, KYC files, claims, invoices. Wherever paper piles up, you have an entry.
• The RPA graveyard: companies that bought automation years ago and quietly gave up. They are pre-sold on the problem and burned on the old solution.
• Warm paths. Shared investors, former colleagues, champions who changed companies. An intro beats a cold open every time.

The output is not a CSV of four thousand domains. Ask for a ranked, reasoned account list and Walle writes one straight into Notion: each account with the evidence it found, the signal that fired, and a specific entry angle. Not "reach out to the VP of Ops" but "they posted three compliance roles since March and are mid ERP migration, lead with the audit-trail story."

Because the reasoning is written down, your team can argue with it. Kill the accounts that miss, tell Walle why, and the next sweep comes back sharper.

Run the hunt weekly. Ask on Monday morning and treat the Notion list as a living document, not a one-off export. Stale lists are how pipelines rot.

And nothing leaves the building without you. Walle researches and writes freely, but any outward action, like emailing a prospect from Gmail, goes through an approval first. You stay the sender. Walle stays the analyst.